J. Agric. Food Chem. 2003, 51, 7879−7883
7879
Antioxidant Activity of Extracts, Condensed Tannin Fractions, and Pure Flavonoids from Phaseolus vulgaris L. Seed Coat Color Genotypes CLIFFORD W. BENINGER*
AND
GEORGE L. HOSFIELD
Sugarbeet and Bean Research Unit, Agricultural Research Service, U.S. Department of Agriculture, Plant and Soil Science Building, Michigan State University, East Lansing, Michigan 48824
Twelve different seed coat color genotypes of Phaseolus vulgaris L. were extracted and pure flavonoids isolated from 10 of these. The seed coat methanol extracts, tannin fractions, and pure flavonoids all displayed antioxidant activity in a fluorescence-based liposome assay. The relatively high activity of the condensed tannin (proanthocyanidin) fractions indicates that these may play an important role in the overall activity of the extracts. This activity also indicates that although these polyphenols cause problems in digestibility, they may be important dietary supplements with beneficial health effects. The pure anthocyanins delphinidin 3-O-glucoside (1), petunidin 3-O-glucoside (2), and malvidin 3-Oglucoside (3) and the flavonol quercetin 3-O-glucoside (4) isolated from seed coats also had significantly higher antioxidant activity than the Fe2+ control. The activity of kaempferol 3-O-glucoside (5) was not different from that of the Fe2+ control. These findings suggest that variously colored dry beans may be an important source of dietary antioxidants. KEYWORDS: Phaseolus vulgaris; antioxidants; extracts; condensed tannin fractions; anthocyanins; flavonols
INTRODUCTION
Dry beans (Phaseolus Vulgaris L.) are widely consumed throughout the world and are an important source of plant protein in many countries (1). In the United States, dry beans have an annual farm gate value in excess of $700 million (2). Considerable variability exists in common beans (P. Vulgaris L.) for seed characteristics, and consumers have acquired specific preferences for various combinations of size, shape, and color of the dry seeds (3). In many countries these preferences distinguish the various commercial (market) classes, which must meet specific consumer expectations and industry standards. Seed coat color and seed size are the two main criteria that identify the numerous market classes recognized throughout the world. Prakken (4, 5) reviewed the literature for the genetics of seed coat color and showed that eight Mendelian loci contribute: P, C, D, J, G, B, V, and Rk. According to Prakken’s interpretation, C, D, and J are the color genes; G, B, and Rk are modifying genes (have an intensifying effect or darkening influence upon pale colors formed by the action of the color genes); and V is called the violet factor. The dominant allele, V, causes bluish or violet to black colors to develop in the seed coat. The seed coat color of dry beans is determined by the presence and amounts of flavonol glycosides, anthocyanins, and * Address correspondence to this author at the Department of Environmental Biology, University of Guelph, Guelph, ON, Canada N1G 2W1 [telephone (519) 824-4120, ext. 54846; fax (519) 837-0442; e-mail
[email protected]].
condensed tannins (proanthocyanidins) (6). Recently, a variety of seed coat color genotypes have been obtained, and the phenolic compounds responsible for color were isolated and identified (7-12). To date, work on P. Vulgaris has focused mainly on the antinutritional effects of seed coat polyphenolics such as condensed tannins (13-15), but nothing has been reported on what beneficial effects, if any, the wide array of phenolic compounds found in P. Vulgaris seed coats may have. For example, polyphenols from dry beans may act as antioxidants to inhibit the formation of damaging free radicals that result from the natural degradation of foods (16). Flavonoids obtained commercially (17-19) and isolated from plant species (20) are known to be effective free radical scavengers. Recently, condensed and hydrolyzable tannins of relatively high molecular weight have also been shown to be effective antioxidants with greater activity than simple phenolics (21). The purpose of the current study was to determine whether P. Vulgaris L. methanolic seed coat extracts exhibited antioxidant activity by inhibiting lipid peroxidation in a fluorescencebased lipid bioassay (22). We also wanted to determine whether isolated condensed tannin fractions and pure flavonoid compounds were responsible, at least in part, for antioxidant activity that may be observed in the methanol fractions. Therefore, fractions containing condensed tannins and pure compounds isolated from the methanolic extract of a variety of color genotypes were also tested to determine their effectiveness as antioxidant compounds.
10.1021/jf0304324 This article not subject to U.S. Copyright. Published 2003 by the American Chemical Society Published on Web 12/03/2003
7880
J. Agric. Food Chem., Vol. 51, No. 27, 2003
Beninger and Hosfield
Table 1. Flavonoid Compounds Present in Seed Coats and Names, Seed Coat Colors, and Abbreviations of the 12 Genetic Stocks Used for
Antioxidant Activity Analysis seed coat color, common name, and (abbreviation) white V0059 gray white (GW) brown buffy citrine (BC) mineral brown (MB) matte mineral brown (MMB) yellow manteca (MT) yellow brown (YB) blue to black 5-593 matte black (MTB) dark brown violet (DBV) red dark red kidney (DRK) light red kidney (LRK)
color genotype
compounds present in seed coat
pgri[C r] d J G b v Rk pgri[C r] D J G B V Rk Asp
condensed tannins* UNK
P [Cr] D J g B v Rk Asp P [C r] D J G B v Rk Asp P [C r] D J G B v Rk asp
K3G, condensed tannins K3G, condensed tannins K3G, condensed tannins
P [C r] d j G b vlae Rk P [C r] D J G b v Rk Asp
K3GX, K3G K3G, condensed tannins
P [C r] D J G B V Rk Asp P [C r] D J G B V Rk asp P [C r] D J G b V Rk Asp
D3G, P3G, M3G, condensed tannins D3G, P3G, M3G, condensed tannins D3G, P3G, M3G, condensed tannins
P [cu] D J g B v rkd P [cu] D J g B v rk
Q3GX, Q3G, K3G, condensed tannins Q3G, condensed tannins
MATERIALS AND METHODS Plant Material. The genetic stocks used for this investigation were developed by M. J. Bassett by backcrossing selected recessive genes for seed coat color into a recurrent parent Florida breeding line 5-593. Line 5-593 has dominant genes for seed coat color in all eight seed coat color loci (P, [C r], D, J, G, B, V, and Rk), except for the dominant red locus R, which is tightly linked to the C locus (23). The genetic stocks are in the third backcross to the recurrent parent 5-593 (24) and are not isogenic with respect to the recurrent parent 5-593, but an isogenic condition is not needed for our experimental purpose. For each seed coat line, the genotype and the compounds it contains are given in Table 1. Genetic stocks were increased in the summer of 1996 in a nursery at the Saginaw Valley Bean and Sugarbeet Research Farm, Saginaw, MI. The soil type on the farm is a Mistequay silty clay [fine, illitic (calcareous), frigid typic Haplaquolls]. After the beans were harvested in the autumn, seeds were frozen at -20 °C. Over several months beans were removed from the freezer and soaked in distilled water, and the seed coats were removed. Seed coats were then frozen in a -80 °C freezer, freeze-dried, and ground to a fine powder. Extraction. One hundred grams of dried (lyophilized), ground seed coats from each genetic stock was loaded into glass columns and extracted sequentially with hexane, ethyl acetate, methanol, and methanol/water (1:1). Extracts were dried under reduced pressure and heat (40 °C) in a rotary evaporator and stored at -20 °C prior to use. Isolation of Condensed Tannin (Proanthocyanidin) Fractions. The crude methanol extract (3 × 30 mg) from each genetic stock was dissolved in 1.0 mL of methanol and then loaded directly onto the surface of a column (14.0 × 1.0 cm i.d.) packed with Sephadex LH20 gel, which had been equilibrated overnight with 100% methanol. Flavonoids, simple phenols, and hydrolyzable tannins were then eluted from the column using methanol/water (8:2) as the eluting solvent. Condensed tannins were selectively retained on the Sephadex gel (25), and these fractions were then eluted using 50% aqueous acetone as a solvent, evaporated to dryness with a rotary evaporator, weighed, and then diluted for use in the antioxidant assay. Isolation and Identification of Individual Flavonoid Compounds. A variety of thin-layer, medium- and high-pressure liquid chromatography techniques were used to isolate the pure flavonoid compounds. Isolated compounds were then identified by 1H and 13C NMR spectroscopy. These methods and identification of these compounds are described in a series of papers (7-12). With the exception of quercetin and kaempferol 3-O-diglycoside, all of the pure flavonoids isolated from bean seed coat genotypes were tested in the antioxidant assay. Each flavonoid was purified three times to give three repetitions for the antioxidant assay. Antioxidant Assay. Preparation of liposomes and the fluorescence spectrographic assay were modified from those of Wang et al. (26)
and Arori and Strasburg (22). A mixture of 157.5 µL of 1-stearoyl-2linoleoyl-sn-glycerol-3-phosphocholine (Avanti Polar Lipids, Inc., Alabaster, AL) and 86.7 µL of the fluorescence probe 3-[4-(6-phenyl)1,3,5-hexatrienyl]phenylpropionic acid (Molecular Probes, Inc., Eugene, OR), diluted 20-fold, was dried under vacuum. The resulting film was suspended in 500 µL of buffer (NaCl, 0.15 M; EDTA, 0.1 mM; MOPS, 10 mM) and then subjected to 10 freeze-thaw cycles in an EtOHdry ice bath. The suspension was then passed 29 times through a polycarbonate membrane with a pore size of 100 nm using a LiposoFast extruder (Avestin, Inc., Ottawa, ON, Canada). Lipid solution was kept out of light during preparation. Prepared liposomes were stored at 4 °C for a maximum of 4 days prior to use. Fluorescence Assay. Liposome solution was kept out of light during the assay. The blank and iron (FeCl2‚4H2O, 0.1 M) control solutions were a mixture of 100 µL of the buffers (HEPES, 50 mM; Tris, 50 mM). Buffers were stored in Chelex 100 to remove metal ions. In the experiments the control solution had 100 µL of buffer, 200 µL of 1 M NaCl, 1.66 mL of N2-sparged Millipore water, 20 µL of DMSO, and 20 µL of liposome solution added. Twenty microliters of iron solution (FeCl2‚4H2O, 0.1 M) was then added to the control solution. The test sample solution was 100 µL of buffer, 200 µL of 1 M NaCl, 1.66 mL of N2-sparged Millipore water, 20 µL of sample solution (1 mg of sample/1 mL of DMSO diluted 10-fold), 20 µL of liposome solution, and 20 µL of iron solution. The final concentration of the samples tested was, therefore, approximately 1.0 µg/mL or 1.0 ppm. Fluorescent intensity of the lipid suspensions was monitored for 21 min using a Turner (Dubuque, IA) model 450 fluorometer. Statistical Analysis. Data were analyzed using the SAS Institute (27) statistical package. Analysis of variance utilized a general linear models procedure (GLM) with a Student-Newman-Keuls comparison of means to test for significant differences between means. RESULTS AND DISCUSSION
Antioxidant Activity of Extracts, Tannin Fractions, and Pure Compounds. Significant differences between the various methanol extracts are given in Table 2 for the readings taken at 15 min into the assay. Red kidney, light red kidney, and the yellow brown extracts were the most active and did not differ from each other significantly. The next most active extracts were those of matte mineral brown, mineral brown, V0059, and buffy citrine, which did not differ significantly in their activity. These genotypes (except for V0059) are varying shades of brown. The next most active extracts were the blue or black series represented by Florida dry bean breeding line 5-593, matte black, and dark brown violet. The least active extracts were those of gray white and manteca. The activity of the crude methanol extracts is due to the presence of flavonoid monomers and polymers (condensed
a V59, V0059; GW, gray white; BC, buffy citrine; MB, mineral brown; MMB, matte mineral brown; YB, yellow brown; MT, manteca; 593, Florida breeding line 5-593; MTB, matte black; DBV, dark brown violet; RK, dark red kidney; LRK, light red kidney. Means followed by a different letter within a row are significantly different at P e 0.05.
0.81a ± 0.01 0.73b ± 0.01 0.53c ± 0.02 0.46d ± 0.01 0.10e ± 0.01 N/A 0.69b ± 0.01 0.44d ± 0.02 0.72b ± 0.03 N/A 0.83a ± 0.01 condensed tannin fractions
0.67b ± 0.02
LRK DRK
0.77a ± 0.03 0.50bc ± 0.02
DBV MTB
0.53bc ± 0.01 0.46c ± 0.01
593 MT
0.17e ± 0.02 0.74a ± 0.01
YB MMB
0.60b ± 0.01 0.60b ± 0.01
MB BC
0.56bc ± 0.01
GW
0.30d ± 0.04
V59
0.58b ± 0.02 methanol extracts
mean relative fluorescencea (1.0 µg/mL)
Table 2. Antioxidant Activity of Methanol Extracts and Condensed Tannin Fractions of Seed Coats of 12 Genetic Stocks Differing in Seed Coat Color
0.70a ± 0.05
Antioxidant Activity of Phaseolus vulgaris Constituents
J. Agric. Food Chem., Vol. 51, No. 27, 2003
7881
tannins), hydrolyzable tannins, and phenolics. Recently, polyphenolic compounds from plants such as condensed and hydrolyzable tannins have been shown to be powerful antioxidants. For example, tannins from Sorghum bicolor Moench were found to be 15-30 times more effective than simple phenols at quenching peroxide radicals (21). Similarly, tannins isolated from adzuki bean (Vigna angularis Ohwi et Ohashi) were also found to have antioxidant properties (28). Finally, tannincontaining extracts from a variety of dry beans (P. Vulgaris) have been shown to inhibit iron-catalyzed oxidation of soybean (Glycine max [L.] Merr.) oil (29). To determine the role the condensed tannin fractions from our seed coat extracts play as antioxidants, condensed tannin fractions were obtained by fractionation of the methanol extracts on Sephadex LH-20 gel as described. When these fractions were tested in the same antioxidant assay used to test the methanol extracts, the condensed tannin fractions were generally found to be as active or slightly more active than the methanol extracts (Table 2). This suggests that a significant amount of the antioxidant activity found in the methanol extracts may be due to the condensed tannins present. However, for matte black, matte mineral brown, and 5-593, the tannin fraction had less activity than methanol extracts. No condensed tannins were detected in the yellow manteca genotype, and its methanol extract (no condensed tannin fraction was obtained) also had the lowest activity of all the extracts tested. This also points to the importance of condensed tannins as responsible for some of the antioxidant activity from methanol extracts. Because individual flavonoids have also been shown to exhibit antioxidant activity (20), these compounds were isolated and identified. Two of the main flavonol glycosides and three anthocyanins were tested in the antioxidant assay. Their structures are given in Figure 1, and their distribution among the different seed coat color genotypes is given in Table 1. Of the beans with white seed coats, V0059 tested positive for the presence of condensed tannins as detemined in the vanillin/ HCl assay. Gray white seed coats did not give a positive response in this assay, and their phenolic composition is not known. Brown and yellow seed coat genotypes had predominantly kaempferol glycosides and tannins with the exception of the manteca type, which did not contain tannins but had an additional kaempferol diglycoside. The blue and black seed coat genotypes all had condensed tannins and contained the same three anthocyanins, but their concentrations differed (12). Finally, dark red and light red kidney had predominantly quercetin glycosides and condensed tannins. The anthocyanins delphinidin 3-O-glucoside and petunidin 3-O-glucoside and the flavonol quercetin 3-O-glucoside were the most active of the pure compounds tested (Figure 2). They differed significantly from the antioxidant activity of butylated hydroxytoluene (BHT) but still inhibited lipid destruction by >50% relative to the control. The third anthocyanin, malvidin 3-O-glucoside, was also active but had significantly less activity than BHT, delphinidin 3-O-glucoside, petunidin 3-O-glucoside, and quercetin 3-O-glucoside. Finally, kaempferol 3-O-glucoside had the least antioxidant activity, inhibiting liposome breakdown by